Project Description
This project is an exploration of spatial data and of different ways to map spatial data in R. Using oil spill data from the California of Fish and Game, I first create an interactive exploratory map in tmap, then create a finalized chloropleth map of inland oil spill events by California county using ggplot().
hide
#Read in the oil spill layer:
ca_oilspill <- read_sf(here("data", "ds394"), layer = "ds394") %>%
clean_names() %>%
rename(name = localecoun)
#Check the projection:
#st_crs(ca_oilspill) #NAD 83 / California Albers, EPSG 3310
#Read in the CA county data shapefile
ca_counties <- read_sf(here("data", "ca_counties"), layer = "CA_Counties_TIGER2016") %>%
clean_names() %>%
select(name)
#Check projection
#st_crs(ca_counties) #WGS 84 / Pseudo-Mercator
#Reset projection
ca_counties <- st_transform(ca_counties, st_crs(ca_oilspill))
#st_crs(ca_counties)
1. Exploratory oil spill location visualization using tmap
Having read in the data and confirmed matching CRS in the code chunk above, here I use tmap to create an exploratory map of all oil spills in California from the data set.
hide
tmap_mode("view") #Set tmap mode to view (for interactive map)
tm_shape(ca_oilspill) + #Make a tmap map of oilspill locations
tm_dots() + #Where dots represent oil spill events
tm_basemap("Esri.WorldTopoMap") #And the base map is ESRI's topomap
Figure 1. Exploratory interactive map of all (inland and marine) oil spill locations in California in 2008.
2. Chloropleth of inland oil spill events in 2008
Now, I wrangle the data set to contain counts of inland oil spill events in 2008 by county. I then plot this data in ggplot() as a chloropleth, using the TIGER shapefiles of California counties to map events by county. Modoc County, which had zero inland oil spill events in 2008, is mapped in gray alongside the chloropleth to indicate that there is no data.
hide
#Wrangling: Want counts of inland oil spill events by county in 2008
modoc <- ca_counties %>%
filter(name == "Modoc") #Make a df with just modoc county for use later
ca_county_oilspill <- ca_counties %>% #Join the two dataframes together
st_join(ca_oilspill)
ca_oilspill_counts <- ca_county_oilspill %>% #Filter for only inland oil spills
filter(inlandmari == "Inland") %>%
group_by(name.y) %>% #Group by county name
summarize(spill_count = n()) #Count oil spill events by county
ggplot(data = ca_oilspill_counts) + #Graph from this count data
geom_sf(aes(fill = spill_count), color = "white", size = 0.1) + #With fill depending on count
scale_fill_viridis() + #Setting the color gradient from viridis package
theme_void() + #With a void theme
labs(fill = "Number of oil spills", #Label for the key
title = "California inland oil spill by county, 2008") + #Label for the map
geom_sf(data = modoc, fill = "lightgray", color = "white", size = 0.1) #Add Modoc county (no spills)
Figure 2. Chloropleth of inland oil spill events in 2008 by California county. Modoc county (gray) had no inland oil spill events in 2008. Los Angeles (yellow) and San Mateo (teal) counties had the two highest counts of inland oil spills (340 and 173, respectively).
Data cited:
California county shapefile data:
“TIGER/Line Shapefiles.” U.S. Census Bureau. Sourced through California Open Data Portal, https://data.ca.gov/dataset/ca-geographic-boundaries
California oil spill data:
“Oil Spill Incident Tracking [ds394]”. California Department of Fish and Game, Office of Spill Prevention and Response. 2008. https://map.dfg.ca.gov/metadata/ds0394.html